带分红交易费和交易税的最优分红问题-概率论与数理统计专业论文.pdf.docxVIP

当初始资金。三x百时，令第一次分红时间和分红额分别为咛=inf{t三Ox;=u}和??二百，第n次分红时间和分红额分别为T=inf{t主Tn一1:x;=u}和巳=U，n22，则最优分红策略就是上面定义下的俨，最优回归函数为V(x)=开[(λ1+μ)(λ3-à2)eà1X+(λ2+μ)(λ1-à3)e).2X+(λ3十μ)(人2-à1)ελ叫，其中Nt∞x;=x+ct一三二写+σW(t)一百汇I(r，在t)τ=1n=1当初始资金z三百时，令第一次分红时间和分红额分别为咛=0和??二x，第n次同上，则最优分红策略就是上面定义下的扩，最优回归函数为V(x)=kx-K，其中N，∞x;=ct-LYi+σW(t)-u2二I(r;;t)i=1口=2关键词:带扰动的古典风险模型;分红;拟变差不等式;最优分红策略;指数索赔分布.AbstractThispaperdealswithoptimaldividendpaymentsintheclassicalriskmodelperturbedbyaBrownianmotion. SinceDeFine吼tti(1957)firstproposedd忖iV叫i诅de口strategiesfori丑nlSU1汀r缸anceriskmodel，dividend problem hasbecome aimportanttopicandoptimaldividendproblemhasbecomeamostpopulartopicintheactuarialliterature.Infact，themainscopeofthispaperistostudytherelation-shipbetweenthesolutionofquasi-variationalinequalityandtheoptimalreturnfunction，thenweobtainthesolutionsoftheoptimalreturn functionandopti-maldividendstrategywhenclaimsareexponentiallydistributed bysolvingthecorresponding quasi-variational inequality.Thisdissertationisdividedintofour sections.Inthefirstsection，westatethebackgroundandthehypothesesoftheopti-maldividendproblem.Inthesecondsectio丑，wegivearigorousmathematicalmodeloftheproblemandseveralkindsofimportant definition.Inthethirdsection，weobtainthecorrespondingquasi-variationalinequalityandprovethe时ationshipbetweenthe optimal return function V(x) and thesolutionof quasi-variational inequality v(x).Ifv(x)iscontinuous on [0，∞)andiscontinuouslydifferentiableon(0，∞)，thenV(x)三υ(x)Intheforthsection，we obtain aclosed-form solution ofquasi-variationalinequalitywhentheclaimssizesareexponentiallydistributedandtaketheopti-malreturnfunctionandoptimaldividendstrategy.Itisshownthattherecanbeessenti?llytwodifferentsolutions:Wheneverthereservereachesabarrier百therearereducedto旦throughadividendpaymentu-1!:，andthereserveprocesscontinues.Optimaldividendstrategyandtheoptimalreturnfuntionareasfollow:Ifthe initial reserve 0 x 百1 the first time and amount of dividendsdefinedby咛=inf{t主0:x;=百}and?r=百一丛then-thtimeandamountofdi